In the medical industry it is well known to use technologies such as computer tomography (CT), positron emission tomography (PET), magnetic resonance imaging (MRI), ultrasound, conventional X-rays and various other technologies to diagnose the health of a patient. The equipment used to provide such imaging is generally very complex and expensive. For CT imaging, the equipment generally includes a CT scanner that collects raw CT data and proprietary software that reconstructs 3D images out of raw data utilizing scanner's various hardware components, including computers. A computed tomography (CT) method uses X-ray scan data to reconstruct detailed images of a body's interior structure.
Despite the fact that only raw CT data preserves all the medical information acquired during a scan, in practice, only the image data (results of reconstruction) are saved. Once raw data are deleted there is no way back to perform additional image reconstructions without a repeat scan. But if raw data are saved, radiologists will be able to request retrospective reconstructions to more precisely zero in on specific regions of interest (ROI) without a repeat scan.
In addition, raw data from previous scans can be used for better planning of new scans and much more accurate monitoring of the treatment/disease progress. Availability of raw scan data to a wider group of image reconstruction professionals will also stimulate faster development and adoption of the next generation of reconstruction technologies. The same data sets can be used for testing and improving other new image reconstruction methods.
The saved raw CT data can also be used to more accurately determine the source of a medical problem, optimize treatment or disease monitoring, and lead to a paradigm shift in medical imaging, improving medical care not just on the individual patient level but on the entire patient population level.
FIG. 1 shows a conventional prior art workflow 1 of CT data acquisition and image reconstruction using software pre-installed on a scanner 20.
Referring to FIG. 1, the Scan Order 10 can be generated by a physician to assist with diagnosing or treating a medical issue or related issues. The scanner technician can perform a scan of a patient in accordance with the Scan Order and certain scanner-specific predetermined scan protocol. The protocol takes into account numerous patient-related variables to optimize diagnostic quality and minimize the amount of radiation exposure by the patient.
During the scan, the scanner collects raw CT scan data, sufficient to perform a computed tomography (CT)-based 3D (three-dimensional) image reconstruction, and stores raw scan data temporarily on the scanner 20 which includes a computer with pre-installed data processing and reconstruction software, and temporary CT data storage.
After the scan is complete, the processing and image reconstruction software preinstalled on the scanner perform raw data processing followed by 3D image reconstruction. The reconstructed 3D image volumes are temporarily stored on the scanner and are also sent to the image repository called Picture Archiving and Communication System (PACS) 30 which also includes a computer for medium-term storage 30.
The PACS system as well as image visualization workstations connected to the PACS can be accessed by physicians/radiologists to read and interpret the reconstructed images of a patient. Long-term storage (e.g., multiple years) of images can also be done.
Prior art exists in the area of image manipulation (or enhancement), which takes place after images have been reconstructed by the scanner system and raw data has been deleted. Such image-based image enhancement has inherent weakness over the raw data-based image reconstruction or raw data-based image enhancement. Many prior art patents in the area of imaging focus on specific algorithms for image manipulation rather than dealing with accumulation and re-inverting (re-imaging) of raw data acquired by a CT scanner’ sensors to reconstruct a more accurate image of the body.
Various patents exist that touch on the idea of using previously generated images or downloading CT image data, but only in terms of using the data for their specific algorithm. See for example, U.S. Pat. No. 7,145,984 to Nishide et al.; U.S. Pat. No. 7,684,589 to Nilsen et al.; U.S. Pat. Nos. 7,599,534 and 7,672,491 to Krishnan et al.; U.S. Pat. No. 7,756,314 to Karau et al.; U.S. Pat. No. 7,860,286 to Wang and Jackson; U.S. Pat. No. 8,195,481 to Backhaus; which are each incorporated by reference in their entirety.
Nishide et al. '984 describes “a method to plan a scan in consideration of past patient exposures. The method includes a step of sampling information on a patient exposure the subject has received during a scan performed for reconstructing tomographic images, which is appended to each of the reconstructed tomographic images, on the basis of identification information with which the subject is identified; a step of creating a distribution of patient exposures calculated relative to an axis orthogonal to the scanning directions on the basis of the sampled information on the patient exposure (an estimated patient exposure, which is estimated in planning a scan, and an exposure limit); and a step of displaying the created exposure distribution.
The components 10, 20 and 30 of FIG. 1 are generally covered in the above identified patents.
Nilsen et al. '589 describes a technique to accelerate the image reconstruction process by dividing one set into subsets, where “The raw image data is decomposed into N subsets of raw image data. N is based on a number of available image generation computer processors. The N subsets of raw image data are processed to create processed image data. The image generation processors perform image processing on the image data in parallel with respect to each other.” This patent describes “a method for increasing the performance of a system for processing raw image data via dividing it into smaller subsets of raw image data. This patent would also be potentially obsolete in the future based on continuing advances in efficiency of algorithms and computing power would make this technology unnecessary.
Krishnan et al. '534 and '491 are described as being used for “processing a medical image to automatically identify the anatomy and view from the medical image and automatically assess the diagnostic quality of the medical image. In one aspect a method for automated decision support for medical imaging includes obtaining image data, extracting feature data from the image data, and automatically performing anatomy identification, view identification and/or determining a diagnostic quality of the image data, using the extracted feature data.” The methods described here are generally using an Image Database for performing feature analysis, anatomy/view identification and quality assessment of the imaged data (images) but not with raw CT data.
Karau et al. '314 is designed for dealing with imaged data for acquiring images on an imaging system and performing accessing image data with a Computer-Aided Design (CAD) algorithm.
Wang et al. '286 describes a medical image acquisition error detection technique which uses special characteristics of medical images to detect possible errors. In general, the presented technique categorizes medical images based on the type of images. When a medical image is to be examined for possible acquisition errors, it is categorized and a measure of difference between the image and the standard image for the category is computed. If the measure of difference falls outside the acceptable difference for the category, the image is deemed to contain an acquisition error and an alert is issued.
Backhaus '481 describes a teleradiology image processing system to receive, process, and transmit radiology read requests and digital radiology image data.
Here, a radiology processing system can include a series of processing components configured to receive digital radiology data from a medical provider, extract relevant information and radiology scan images from the digital radiology data, and initiate and control a workflow with a qualified remote radiologist who ultimately performs a read of the radiology scan images.
Other techniques in this patent facilitate data processing within the image processing system in response to medical facility rules and preferences; translation or conversion of digital images to other formats; compilation of patient and medical facility data obtained from the digital radiology data into medical records or data stores; assignment of radiology studies within a teleradiology workflow in response to licensing and credentialing rules; and billing functions in response to completed reads by the remote radiologist.
Thus, the first key problem in the prior art workflows, including solutions currently utilized by the medical industry, is that the raw data are only temporarily stored on the scanner (i.e., a few days) or even deleted right after the image reconstruction process is finished.
Another problem of the current workflow is that the scanner operates like a “black box”, where no one except the scanner manufacturer has access to the raw scan data. As such, CT scanners are designed in such a way that it is impossible for anyone except the scanner manufacturer to take the raw data from the scanner to perform an additional image reconstruction with another, potentially superior or more customized, image reconstruction algorithm (software).
It is also impossible for third parties to independently install, potentially superior or more customized, image reconstruction software onto the scanner.
When there is a medical diagnosis needed to further scrutinize a particular target using computed tomography, because raw CT data has been deleted, there is no known reliable alternative which is not based on manipulation with images, but to repeat a CT scan, thereby exposing the patient to additional high doses of X-ray radiation. Not only is this re-image requirement detrimental to the health of the patient, but it requires additional costly system resources, medical personnel time, etc.
Moreover, a repeated scan is often performed on a different scanner, and the image reconstruction is often performed by a different software than the one used after the first scan, and the reconstructed images are interpreted by a different person such as a radiologist with different skills, experience, etc.
All these factors can accumulate and result in an inconsistency with the first result and even in an incorrect medical diagnosis. It would be much safer for a patient and more reliable from the medical point of view to resolve a possible issue by returning back to the saved raw data and perform CT reconstruction avoiding all or the majority of the factors mentioned above. On top of that, it would be greatly beneficial to the patients to have an option to perform a repeat reconstruction (of course, without a repeat scan) using the best available reconstruction software. This option is not available in the current art.
These problems not only result in radiation over-exposure of the patient population, but on a global level limit the full potential of medical imaging diagnostic quality and the pace of making available for doctors and patients novel imaging algorithms developed outside the walls of scanner manufacturer technology centers.
Solutions to the above problems are not contemplated by the scanner manufacturers or in the prior art.
Thus, the need exists for a new workflow and system that solve the above problems with the prior art.